The background description provided herein is for the purpose of generally presenting the context of the disclosure. Unless otherwise indicated herein, the materials described in this section are not prior art to the claims in this application and are not admitted to be prior art by inclusion in this section.
Face detection is widely employed in many applications. Most face detection techniques involve a large number of random accesses of a look-up table (LUT) that maps a pattern of its feature to a score value for the pattern to represent a part of the face. Due to the complexity of face detection, the LUT is often too big to be fully fitted into the cache memory of many client computing devices, such as smartphones, computing tablets, and so forth. Thus, on these client computing devices, cache misses occur frequently on the devices during face detection, and as a result, face detection is often not completed as fast as it is desired.